View - ResearchGate

30 Socially Intelligent Agentsto develop joint plans and to carry out coordinated action, and (v) the ability toform persistent relationships and shared memories with other individuals.There is some systematic psychological research on the dynamics of closerelationships, establishing for example their connection with attachment [5].Although knowledge-based cognitive approaches have been used for describingdiscourse, there has not yet been much extension to describing relationships [6].Presumably, a socially intelligent agent would recognize you to be a person,and assign a unique identity to you. It would remember you and develop detailedknowledge of your interaction history, what your preferences are, what yourgoals are, and what you know. This detailed knowledge would be reflected inyour interactions and actions. It would understand and comply with prevailingsocial norms and beliefs. You would be able to negotiate shared commitmentswith the agent which would constrain present action, future planning and interpretationof past events. You would be able to develop joint plans with the agent,which would take into account your shared knowledge and commitments. Youwould be able to act socially, carrying out coordinated joint plans together withthe agent.We would also expect that joint action together with the agent would proceedin a flexible harmonious way with shared control. No single agent would alwaysbe in control, in fact, action would be in some sense voluntary for all participantsat all times.To develop concepts and computational mechanisms for all of these aspectsof social relationship among agents is a substantial project. In this paper, we willconfine ourselves to a discussion of joint planning and action as componentsof social behavior among agents. We will define what voluntary action mightbe for interacting agents, and how shared control may be organized. We willconclude that in coordinated social action, agents voluntarily maintain a regimeof mutual control, and we will show how our agent architecture provides theseaspects of social relationship.2. Our Agent ArchitectureIn this section we describe of an agent architecture that we have designedand implemented [2] [3] and which is inspired by the primate brain. The overallbehavioral desiderata were for an agent architecture for real-time control of anagent in a 3D spatial environment, where we were interested in providing fromthe start for joint, coordinated, social behavior of a set of interacting agents.Data types, processing modules and connections. Our architecture is a setof processing modules which run in parallel and intercommunicate. We diagramtwo interacting agents in the figure. This is a totally distributed architecture withno global control or global data. Each module is specialized to process onlydata of certain datatypes specific to that module. Modules are connected by a